If you add the ML.Net package version 0.0.0.4 to the project, trainers Like OrdinaryLeastSquaresRegressor, and LightGbmRegressor will be visible, and you can add them to the pipeline.
If you run a pipeline with those trainers, and just the ML.Net package on the project, you'll get a runtime exception, with message like:
System.InvalidOperationException: Entry point 'Trainers.OrdinaryLeastSquaresRegressor' not found at Microsoft.ML.Runtime.EntryPoints.EntryPointNode
This is because the actual implementation for those learners lives in additional packages, like Microsoft.ML.LightGBM, or Microsoft.ML.HadLearners.
To get those trainers to work with the pipeline, one needs to:
1- Add the respective package to the project
2- Add the following like to their section in the .csproj file
<CopyLocalLockFileAssemblies>true</CopyLocalLockFileAssemblies>
This will gather all the dlls in the bin folder, and everything should work correctly.
If you add the ML.Net package version 0.0.0.4 to the project, trainers Like
OrdinaryLeastSquaresRegressor, andLightGbmRegressorwill be visible, and you can add them to the pipeline.If you run a pipeline with those trainers, and just the ML.Net package on the project, you'll get a runtime exception, with message like:
System.InvalidOperationException: Entry point 'Trainers.OrdinaryLeastSquaresRegressor' not found at Microsoft.ML.Runtime.EntryPoints.EntryPointNodeThis is because the actual implementation for those learners lives in additional packages, like Microsoft.ML.LightGBM, or Microsoft.ML.HadLearners.
To get those trainers to work with the pipeline, one needs to:
1- Add the respective package to the project
2- Add the following like to their section in the .csproj file
<CopyLocalLockFileAssemblies>true</CopyLocalLockFileAssemblies>This will gather all the dlls in the bin folder, and everything should work correctly.